Reactive teaching and informed feedback

Because TBL, along with Dogme and CLL, involve reactive teaching, language needs are addressed both during and after the task itself. Practitioners of task-based learning have given it a rather technical name, Focus On Form.  This simply means that language focus occurs in the context of communication, so that the lesson itself remains centred on task completion, and what students can do with language, rather than on language items for their own sake. This isn’t just a semantic distinction – it underpins the whole approach. Teachers who follow any of these fluency-led approaches tend to believe that:

“form can best be learned when the learner’s attention is focused on meaning” (Beretta, 1989, quoted in Thornbury’s Walking While Chewing Gum article.)

Most writers now agree on its effectiveness; few hold firmly to a so-called non-interface position. Ideally, this kind of corrective feedback can be delivered at the exact point at which learners need it. 

One of the advantages of this approach – language focus as corrective feedback rather than a conventional clarification stage in lessons – is that teachers can cast their net wide, to cover emergent language, including structures, functional language, lexis (individual words, collocations and formulaic language), and pronunciation, none of which requires intensive practice. These can form the kind of varied language diet which is likely to be more nutritious for students than Thornbury’s infamous “grammar McNuggets”. 

One of the things that prevents teachers being supplanted by AIs is that (for the foreseeable future at least) only human teachers can provide feedback that is targeted towards learners’ actual needs (i.e. they can perceive what is merely a slip and what is a developmental error that may respond better to correction); they can take into account factors such as a learner’s level, personality type and receptivity to correction; and they can provide additional value by introducing more advanced or idiomatic language to help students say something better (sometimes called “upgrading” or “feeding forward”.)

Technology has added to the ways that we can provide this kind of feedback: 

  • in the chat box
  • on a Google document
  • through (online or smartphone) recordings.
Photo by Black ice on

“both immediate and delayed correction may be beneficial”

Ellis & Shintani

Which language?

During tasks, teachers are likely to encounter a host of instances of students’ output to which they would like to respond. To be most efficacious and to make feedback “stick” which examples should be chosen? Given we can’t see inside students’ heads, it’s a very subjective matter and there is with little guidance available.

Ellis (2009:6) recommends focusing on a few targeted errors rather than having a smorgasbord of correction, which may be inhibiting. Global errors are likely to seem more important than local errors because they result in a garbled message. I remember when my French teacher usefully chose to target my (non-) use of the subjunctive; this has stuck with me because it was an instance of what I wasn’t doing in my spoken language. Consistent errors like these – including such errors of omission – look as if they’ll respond well to correction as they are indicators of a student’s developmental stage, whereas mere slips don’t seem worth bothering with if the student “knows” the item, at least explicitly.

Kerr (2020:6) recommends strongly that teachers do not stick to feedback on grammatical accuracy, but expand their feedback to cover all aspects of a task’s performance.

Larsen-Freeman (2003:133) writes that the most important errors to correct are ones about which the student won’t otherwise receive any evidence from the natural input available in their environment; of course these are not always easy to determine. There will be more cases of these in EFL (rather than ESL) environments because of the amount of exposure students get to English, so Doughty and Williams (1998:200) argue that the learning context is therefore one of the most important factors to consider in determining one’s approach to the focus on form, and of course this is partly a socio-political decision, based on which variety of English is spoken in that context. 


Teachers are often told on training courses that feedback should be delayed in fluency-based activities, but immediate during some accuracy-focused “controlled practice”. In reality, the distinction is not as clear cut. Writers from Larsen-Freeman to Thornbury agree with the effectiveness of on-the-spot correction. Doughty in particular has argued in favour of immediate correction to take advantage of a window of opportunity for the student to make the connection between form and meaning, but Lethaby et al (2021:201) say that this contention – that there is some critical window within which feedback must take place in order to be effective – is “not supported by research.” While there are no firm (universalisable) conclusions about when correction should occur, the common idea that correction will disrupt fluency is unlikely to be correct, according to Ellis & Shintani (2014:280), who state that, “both immediate and delayed correction may be beneficial.”.


Teachers can do a combination of any of the following:

  • Indicate that an error has occurred, and where –  to help the student self-correct
  • Provide a reformulated model 
  • Provide metalinguistic information 

Let’s look at the second of these; this is one of the most common kinds of feedback in classrooms, though not without its problems.

Recasts v explicit correction

Reformulation is when teachers provide a correct model of what the student wanted to say – the most natural kind of response to students’ language – and recasts are a subset of this. Here is an example:

S: It’s not so much cold.

T: Yes, you’re right. It isn’t all that cold today.

Recasts are implicit: while they provide information, they don’t have the affective force of correction, appearing as simply part of the natural conversation. The disadvantage, of course, is that a student who is not fully alert may not notice the correction unless made to repeat the corrected version, or unless the teacher captures the language on the whiteboard later.

Recasts, which have some heavyweight defenders including Doughty and Long, have several advantages:

  • they are fully contextualised; the student’s meaning is clear
  • they come soon after the error so that the student can “notice the gap” between their own production and the correct model
  • the student’s attentional resources are available to focus on the form
  • they are brief.

Doughty et al argue that it’s possible that explicit correction may be more effective than a recast where there are simple pedagogical “rules” that we can give students. (1998: 232) Some research shows that students prefer explicit correction.


Students can over-use certain forms or be unable to find the most idiomatic way to say something. This is where an “upgrade” can work well. Here are two examples:

S: I watched 15 episodes on Netfix last night.

T: Really? You binge-watched 15 episodes?

S: Yesterday I was to.. to go the main office….because the manager, of the region he must to come..…..but he didn’t arrive, so I don´t go and I had a free day finally. It was a fantastic!

T: You were supposed to go, but in the end you didn’t have to. Great! 


In our book, Activities For Task-Based Learning, we offer a framework for dealing with emergent language: ICE (identifying – capturing – exploiting). We’ve also collected some practical ways to help vary feedback. We suggest using various feedback templates, both for groups and individuals. Here is one example:

Teachers can note some student errors on slips of paper, which are given to the student on the spot and can be reflected on later; these are especially useful in the case of early finishers. The whiteboard can obviously be used to record examples for delayed correction. Google documents is an easy tool to use in online classes, and is particularly effective in writing lessons, as groups or individuals can write in different parts of the document and the teacher can highlight sections which need reformulating, or use the comments function.

Some conclusions

According to Larsen Freeman (2003:126), language is a dynamic system which responds to feedback, and that as such, “providing feedback is an essential function of teaching”. 

Despite the benefits, “applied linguists now generally agree that we are unlikely ever to be able to identify the perfect recipe for giving feedback to language learners.” (Kerr, 2020:21). 

It appears that students benefit in particular from negative evidence about grammar or vocabulary use i.e. where they are trying something out and they need to be told that a particular word combination or structure doesn’t work. This negative evidence is sometimes not available in the natural input, and so teachers can usefully focus their correction here, helping with persistent, fossilised errors and “false friends” where the L1 provides a misleading cue. As well as simply correcting, teachers can help students communicate more effectively by “upgrading” and offering more idiomatic lexis; giving explicit information about grammatical forms where the language point  appears to be difficult; and helping with task performance in other ways, for example speaking skills. Feedback should be selective. Recasts should ideally be kept brief, whereas delayed correction creates the opportunity to explore a broader range of language. 

Task repetition then provides an opportunity to try out some of the new language – and this has been shown to result in gains in accuracy, fluency and complexity.


Doughty & Williams (eds.) (1998) Focus on Form in Classroom Second Language Acquisition, Cambridge Applied Linguistics

Ellis (2009) Corrective Feedback and Teacher Development in L2 Journal 1 (1)

Ellis & Shintani (2014) Exploring Language Pedagogy through Second Language Acquisition Research, Routledge

Kerr, P. (2020) Giving feedback to language learners. Part of the Cambridge papers in ELT series: Cambridge University Press

Larsen-Freeman, D. (2003) Teaching Language: From Grammar To Grammaring, Heinle

Lethaby, Mayne and Harries (2021) An Introduction to Evidence-Based Teaching in the Language Classroom, Pavilion.

Long, Mike (2015) Second Language Acquisition and Task-Based Language Teaching, Wiley.

Meade-Flynn, Emma (2017) 10 ways of moving from feeding back to feeding forward, conference presentation, IH Barcelona

Thornbury, Scott (2000) Walking While Chewing Gum, IH Journal